Spatial spillover effects of transport infrastruct

Spatialspillovereffectsoftransportinfrastructure:evidencefromChineseregionsNannanYua,b,⇑,MartindeJonga,b,ServaasStormb,JianingMiaaHarbinInstituteofTechnology,WestDazhiStreet92,150001Harbin,ChinabDelftUniversityofTechnology,Jaffalaan5,2628BXDelft,TheNetherlandsarticleinfoKeywords:TransportinfrastructureSpillovereffectsEconomicgrowthSpatialDurbinModelChinaabstractThispaperexaminesthepossibilityofspatialspillovereffectsoftransportinfrastructureinChineseregions.WeestimatetheregionalspilloversofthetransportinfrastructurestockbyapplyingaspatialDurbinModelforthetime-period1978–2009,andalsothreesub-periods,1978–1990,1991–2000and2001–2009.Theresultsindicatethatpositivespilloversexistineachperiodduetotheconnectivitycharacteristicoftransportinfrastructureatthenationallevel.Attheregionallevel,transportinfra-structurespillovereffectsvaryconsiderablyovertimeamongChina’sfourmacro-regions:theeasternregionenjoyedpositivespilloversallthetime;thenortheasternregionhadnosignificantspillovereffectsin1978–1990,negativespilloversin1991–2000,andpositivespilloversin2001–2009;thecentralregionhadnegativespilloversforthethreesub-periods;forthewesternregion,negativespilloverscanbeobservedafterthe1990s.TheanalysisindicatesthatchangesinspilloversamongregionsarecloselyassociatedwiththemigrationofproductionfactorsinChinaduringthelastdecades.2012ElsevierLtd.Allrightsreserved.1.IntroductionPlentyofstudieshavebeenconductedontheimpactoftrans-portinfrastructuredevelopmentonregionaleconomicgrowthoverthelastdecades,mostlyaimingtoexaminetheeconomicreturnsoftransportinvestmentsinordertofindareasonableinvestmentpattern(Aschauer,1989;Munnell,1990;Ozbayetal.,2003;Can-ningandBennathan,2007).Eventhoughtherangeofthemeasuredeconomicgrowtheffectsvarieswidelyamongstudies,thepositiverelationshipbetweentransportinvestmentandeconomicdevelop-mentisnowcommonlyaccepted(BanisterandBerechman,2001;Berechmanetal.,2006).Howeverthefindingthattheimpactoftransportinfrastructureattheregionallevelisgenerallylowerthantheresultsobservedatthenationallevelleadssomeresearch-erstoconcludethatthereexistsignificantspillovereffectsacrossregions.Subsequentresearchhastriedtoconfirmthis(Munnell,1992;Holtz-Eakin,1994;CohenandPaul,2004;Cantosetal.,2005;Berechmanetal.,2006;Ozbayetal.,2007).Attemptshavebeenmadetocorroboratetheclaimthatthepositivebenefitsaccruingfromtheseinvestmentsderivenotonlyfrominvestmentsmadebyindividualstates,butthattherearealsopositiveexternal-itiesfromnetworkexpendituresmadebyneighboringstates(Lall,2007).Thatisbecausesomeeffectsinducedbytransportinfra-structurewillextendoutsidethelimitsofthisarea,generatingspillovereffects(Munnell,1992;Boarnet,1995,1996,1998).Onlyafewofstudieshavefocusedonascertainingthepossibleexistenceofregionalspilloversfromtransportcapital,probablybe-causeitisdifficulttofindcasesofcountriesinwhichitsterritoryisdividedinregionswithsubstantialpoliticalpowerastheUSAandSpain(Cantosetal.,2005).ForthecaseoftheUSA,ontheonehand,Munnell(1992)foundthattheimpactsofhighwaycapitalbecamesmallerasthegeographicfocusnarrowedThusshehypothesizedthathighwaypubliccapitalcancreatepositivecross-statespill-oversbecauseofproductivityleakages(spillovers),sincethetrans-portinfrastructurehasnetworkcharacteristics.ButHoltz-EakinandSchwartz(1995)rejectedthisargumentaftermeasuringthespillovereffectsseparately.Ontheotherhand,Boarnet(1995,1996)hypothesizedthatpubliccapitalinfluenceseconomicactiv-ityIargelybyshiftingthatactivityfromonelocationtoanother,andseesthisclaimconfirmedinthecaseoftheUSstreet-and-highwaycapital.Consideringthesetwoarguments,Berechmanetal.(2006)investigatedthespilloversoftransportationatthestate,countyandmunicipalitylevelsoftheUSArespectively,andtheyconcludedthatthespilloversexistatthesmallgeographicareas(atthemunicipalitylevel)butthattheycannotbefoundatthestateandcountylevels.Ozbayetal.(2007)calculatedthecon-tributionoftransportinvestmentstocountyoutputusingthedatafromtheNewYork/NewJerseymetropolitanarea,andtheirresultsshowedthatthespillovereffectsdecreasedwithdistancefromthe0966-6923/$-seefrontmatter2012ElsevierLtd.Allrightsreserved.⇑Correspondingauthorat:HarbinInstituteofTechnology,WestDazhiStreet92,150001Harbin,China.Tel.:[phone].E-mailaddresses:[e-mail](N.Yu),[e-mail](M.deJong),[e-mail](S.Storm),[e-mail](J.Mi).JournalofTransportGeography28(2013)56–66ContentslistsavailableatSciVerseScienceDirectJournalofTransportGeographyjournalhomepage:findingsshowthevariance:Cantosetal.(2005)confirmedtheexistenceofverysubstantialspillovers;Álvarezetal.(2006)didnotfindevi-denceoftheexistenceofspillovereffectsofpublicinfrastructureusingapaneldatasetofthe47mainlandSpanishprovinces;MorenoandLopez-Bazo(2007)investigatedtheeconomicreturnstolocalandtransportinfrastructureandfoundnegativespilloversacrossSpanishregionsintransportcapitalinvestments.Thisvarianceoftheresultsisprobablybecauseofthedifferencebetweens